Mistakes

Costly Mistakes in Motor and Generator Manufacturing (and How to Catch Them)

A troubleshooting guide to the winding, balancing, potting, and test failures that quietly wreck motor and generator yield, and the specific checks that catch each one.

The most expensive early mistake is trusting a nameplate copper fill instead of measuring it. Symptom: your Stator Winding Labor estimate holds but slots jam on insertion and end-turn height runs 3 to 4 mm over drawing. Root cause is a fill factor computed from bare conductor area while the shop wound enameled wire, so a stated 0.42 slot fill is really 0.48 after insulation and varnish swell. Fix: pull grade 2 enamel build (roughly 0.06 to 0.09 mm per side on 1.0 mm wire) and slot liner thickness into the Copper Fill Factor calculator before release, and cap useful fill at 0.45 for hand insertion, 0.55 for needle winding.

Unit slips on winding labor bleed straight into the quote. Symptom: standard time per stator looks fine at 18 minutes but the line misses takt by 30 percent. Root cause is mixing turns per coil with turns per pole, or counting a 3 phase, 4 pole, 12 slot pattern as 12 coils when concentric winding gives 6. Fix: reconcile total conductor length against measured wire draw. If Stator Winding Labor assumes 42 m per stator and the spool log shows 63 m consumed per unit, your turn count or mean length of turn is wrong by 50 percent, and the labor minutes are too.

Rotor balancing quietly eats machine time when the correction plane and grade are mismatched. Symptom: two plane balancing runs stretch to 6 or 7 iterations instead of 2, and Rotor Balancing Time overshoots by triple. Root cause is chasing a G1.0 grade on a fan-cooled rotor that only needs G2.5, or applying single plane logic to a rotor whose length to diameter ratio exceeds 0.5. Fix: set residual unbalance from ISO 21940 (G2.5 at 3000 rpm on a 12 kg rotor allows about 95 g mm per plane), and expect convergence in 2 to 3 runs once influence coefficients are recalibrated after any tooling change.

Hipot rework is usually a test setup problem masquerading as a build defect. Symptom: Rework from Failed Hipot spikes to 8 percent while insulation resistance readings look healthy. Root cause is ramp rate and voltage set too aggressively, or leakage current threshold set at 0.5 mA when winding capacitance alone draws 0.8 mA at 1500 V AC. Fix: switch borderline AC hipot to DC to remove capacitive charging current, ramp at 500 V per second, and set the trip from measured population scatter, not a round number. A properly set threshold should drop nuisance failures below 1 percent and keep true escapes near zero.

Skipping cure verification on varnish turns a load calculation into a scrap generator. Symptom: Insulation Varnish Cure Load says the oven holds 40 stators per batch, but units come out tacky and fail dielectric a week later. Root cause is loading to volume, not thermal mass, so a 40 unit batch draws the 180 C setpoint down and starves the center of the rack of the 2 to 3 hours at temperature it needs. Fix: derate the batch by measured recovery time, verify a through-cure with a thermocouple on a center part reaching setpoint within 45 minutes, and treat oven capacity as thermal, not geometric.

Magnet cost blows up when the bill of materials ignores grade, coating, and scrap. Symptom: Magnet Cost per Motor lands 20 to 35 percent under actual. Root cause is quoting N42 at last year's neodymium price while the design needs N48H for a 150 C rotor, and omitting the 3 to 6 percent handling breakage on thin arc segments. Fix: price the actual grade with its dysprosium loading, add nickel-copper-nickel coating cost, and carry a breakage allowance. On a 6 magnet rotor at 14 grams each, a 5 percent chip rate plus grade correction can move landed magnet cost from 4.10 to 5.60 dollars per motor.

End of line test stations get under-provisioned because takt and test time are confused. Symptom: Generator Assembly Takt shows 90 seconds per unit but the End-Of-Line Electrical Test bench needs 210 seconds for surge, hipot, and no-load spin. Root cause is treating a serial test as if it fits inside takt with one fixture. Fix: size stations by test time divided by takt, so 210 over 90 means 3 parallel benches plus spares, not one. Under-sizing here is why a line that hits assembly takt still ships late, with 40 percent of units waiting at test.

Winding scrap is under-reported because rejects are valued at scrap copper, not built value. Symptom: Winding Scrap Cost reads a few dollars per reject while the P&L shows a much larger hole. Root cause is booking a failed stator at the 8 dollar copper reclaim value when it already absorbed 22 minutes of labor, varnish, liners, and oven time. Fix: value each scrapped stator at full accumulated cost at the point of failure, so a unit that fails final hipot after cure is a 40 to 55 dollar loss, not an 8 dollar one. That correction usually reprioritizes which defect to chase first.

Published 2026-07-01.